name: MiniMax-M2.7-REAP-172B-A10B-MLX-4bit
base_model: MiniMaxAI/MiniMax-M2.7
license: other
pipeline_tag: text-generation
tasks: text-generation
language: en
library_name: mlx
tags:
- Cerebras
- MiniMaxAI
- M2.7
- REAP
- MLX
- static quantization
- 4-bit

Description

This is a 230 billion parameter MiniMax M2.7 model with 25% of its experts pruned with REAP (Router-weighted Expert Activation Pruning), then converted to MLX with mlx_lm.

and sequence using source version of mlx_lm from source and mlx:

hf download saricles/MiniMax-M2.7-REAP-172B-A10B-BF16 --local-dir MiniMax-M2.7-REAP-172B-A10B-BF16
mlx_lm.convert --hf-path saricles/MiniMax-M2.7-REAP-172B-A10B-BF16 --mlx-path ~/Downloads/MiniMax-M2.7-REAP-172B-A10B-MLX-4bit -q --q-bits 4

exdysa/MiniMax-M2.7-REAP-172B-A10B-MLX-4bit

Description

exdysa/
MiniMax-M2.7-REAP-172B-A10B-MLX-4bit